Translation and Validation of the Digital Working Alliance Inventory (DWAI) into German

The Digital Working Alliance Inventory (DWAI) is a measure designed to evaluate how users connect with digital health tools. This project focuses on translating the DWAI into German and validating its use in unguided digital health interventions by combining conventional translation methods with modern AI-supported procedures.

Digital health interventions have become increasingly important because they offer flexible, accessible, and low-cost support for different target groups. These interventions can be categorized as guided (with human support via chat, email, or video) or unguided (fully automated without human involvement). Unguided interventions are particularly attractive due to their immediate availability and scalability.

A key factor that contributes to the success of these interventions is the working alliance, which describes the relationship users develop with the digital tool. Research has shown that a strong working alliance can positively influence health outcomes. However, most existing working alliance measures were originally developed for face-to-face therapy and may not fully capture the unique features of unguided digital tools.

The DWAI was developed by Henson and colleagues in 2019 to specifically assess the working alliance in unguided digital health interventions. It includes six items covering three important components: Goals, tasks, and bond. While the DWAI shows strong psychometric quality in English-speaking populations, a German version had not previously been available. This limits its application in German-speaking settings and makes cross-country comparisons more difficult. Translating and validating the DWAI helps ensure consistent assessment of working alliance in digital health.

The emergence of artificial intelligence (AI) technologies, particularly Large Language Models, presents new opportunities for translation procedures that may be more time- or resource-efficient. However, studies evaluating AI-supported translation versus conventional translation procedures and quality are limited. This project explores different translation procedures from English to German to evaluate how AI could complement the translation of validated measures.

The DWAI-G is currently being validated in a three-arm randomized controlled trial involving patients with cancer-related fatigue or sleep problems. Participants complete a 6-week unguided online course in acupressure or mindfulness-based techniques. The DWAI-G is administered after week 1, with a retest in week 2. The study examines structural validity, internal consistency, test-retest reliability, and construct validity using preregistered hypotheses. The results will show whether the DWAI-G reliably measures working alliance in German-speaking users of unguided digital health interventions.

Krebsforschung Schweiz KFS-5723-02-2023

Claudia Witt (Principal Investigator)

Anna Katharina Tietjen (Contact)